Skill Map
The FDE Skill Map
FDEs need enough depth to build, enough product sense to choose the right problem, and enough field judgment to make the system survive contact with reality.
Core capabilities
Model behavior, prompting, tool use, structured output, context windows, latency, and cost tradeoffs.
Connecting SaaS systems, internal services, identity, permissions, queues, and workflow tools.
Retrieval, data freshness, access control, auditability, and source-of-truth discipline.
Task decomposition, tool calling, review steps, human approval, retries, and rollback paths.
Quality metrics, regression checks, traces, failure analysis, and production feedback loops.
Discovery, process mapping, stakeholder alignment, training, and adoption management.
What separates strong FDEs
A strong FDE is not only fast at implementation. They can decide which AI workflow is worth building, spot where the operational risks live, and shape the deployment so the customer can actually maintain it.
FDE 的真正价值不是写几个 prompt，而是把 AI 变成组织可以依赖的生产能力。